NEURO NEWS

Scientists in the US have developed an artificial intelligence (AI) powered algorithm to analyse video recordings of clinical sleep tests, that could help with early detection of dementia or Parkinson's disease.

Research has long suggested that sufferers of a condition known as REM sleep behavior disorder (RBD) - which sees people act out their dreams - have increased inflammation in an area of the brain where the chemical dopamine is produced.

Parkinson's and dementia patients are known to have diminished supplies of dopamine because nerve cells that make it have died. Experts believe that most people with RBD could very likely be exhibiting early signs of dementia or Parkinson's disease.

RBD affects around 5% of people and includes talking, laughing, shouting and swearing while sleeping. Most people with this disorder thrash around in bed and feel bleary-eyed the next day, often falling asleep during the daytime. In some cases their nocturnal movements are so vigorous that they hurt themselves or their partners.  

The condition is extremely difficult to diagnose because its symptoms can go unnoticed or be confused with other diseases. A definitive diagnosis requires a sleep study, known as a video polysomnogram, to be conducted by a medical professional at a facility with sleep-monitoring technology. 

A recent study, carried out at Mount Sinai School of Medicine in New York, used 2D cameras to assess the clinical sleep tests of 170 patients, 80 of whom had RBD. The other 90 either had another sleep disorder or no sleep disruption. 

The automated algorithm tracked the motion of pixels in the video generated by sleep movements during REM sleep. Researchers then used data to extract the rate, ratio, magnitude, and velocity of movements, and ratio of immobility. 

The technology demonstrated an accuracy rate of 92% in identifying RBD - the highest known result to date - which could prove extremely useful in also pinpointing those at higher risk of developing dementia and Parkinson's disease.